Thirty three University of Limerick students were asked to participate in a research concerning implicit and explicit attitudes toward smoking-related social identities, for example smokers and non-smokers. Implicit Relational Assessement Procedure (IRAP) was enlisted in this research to understand an individual’s implicit attitudes. Past research has indicated that we are often unaware of these attitudes as self presentation strategies mediate implicit attitudes if they are socially undesirable. A 2 x 2 mixed design was entailed, the DV’s were response latencies on the IRAP and attitude responses on the Explicit’s scale. The research asked participants to indicate their smoking history (which was an IV) and they had to respond to a computerised task both in a consistent and inconsistent manner, depending on their behavioural history (the second IV). There was two practise blocks and six test blocks, participants had to score over eighty per cent and respond to the stimuli within two seconds. IRAP eliminates social buffering and explicit attitudes were gathered in an attitude thermometer scale.

The main findings indicate that smokers respond to smokers and non-smokers positively both implicitly and explicitly. Non-smokers respond implicitly and explicitly in a more positive manner toward non-smokers. However non-smokers are implicitly ambiguous toward smokers, yet explicitly they condone smokers. There is an interaction effect in the Explicit scale as non-smokers score higher in being anti-smoker than pro non-smoker.

Introduction

Within society, smoking has always been a topical issue which has experienced significant attitude change in the last fifty years. In the past, smoking was considered a positive social behaviour; however research linking smoking to throat and lung cancer has led to negative attitudes toward smoking and smokers. As smoking is considered detrimental to one’s health, the Irish government have implemented a number of anti-smoker strategies, for example advertisement campaigns on the negative effects of smoking, high taxes on cigarettes.

However the most influential of these was the smoking ban which was sanctioned in 2003. This law prohibits smoking in public locations, where smokers are forced outside consequently. It is now of the social opinion that it is “rude to light up without asking the permission of those sharing your breathing space” (Euranet, 2010).

This research elaborates from the study of Vahey et al who investigated smoking-related social identity preferences within adolescents; their findings indicated that adolescent smokers are more likely to hold positive regard for smokers, in comparison to non-smokers. However non-smokers impression of smokers and non-smokers did not differ, this may be due to peer pressure i.e. striving to fit in.

To expand from this study, our research will be carried out in the adult population as peer pressure is less evident at this stage. Within this research it is hoped to reveal both implicit and explicit attitudes toward smoking-related social identity. Implicit attitudes relate to “hidden nature of social attitudes” which are not prevalent explicitly due to “self presentation strategies,” (). In order to uncover these implicit attitudes, the Implicit Relational Assessment Procedure (IRAP) was used, the hypothesis of the experiment states: Participants should have faster response latencies in consistent rather than inconsistent measures. Shorter response times indicate that the stimuli are consistent with one’s attitude. IRAP presents inconsistent stimuli, it is expected that participants will take longer to respond as this does not correspond with their views which eliminates social buffering.

The research hypothesis affirms that non-smokers will both implicitly and explicitly regard themselves positively as being a non-smoker is considered socially desirable. Non-smoker’s will be more negative toward smokers implicitly and explicitly due to social attitudes, developed through the introduction of the smoking ban, adverts etc. Smokers will regard themselves positively as smoking is considered part of their social identity but also deem non-smokers positively but not as highly as non-smokers. There will be no interaction effect between smokers and non-smokers implicitly and explicitly.

Method

Participants

The sample consisted of 33 undergraduate students from University of Limerick; including eight males and twenty five females. There were 18 smokers and 15non-smokers, all of whom participated voluntarily.

Materials

A questionnaire containing demographics, identified their smoking status was employed, this also served as a primer as it asked about different aspects of smoking for example “have you ever lived with anyone that smoked?”

IRAP was used to understand the implicit attitudes associated with smokers and non-smokers. This is a computerised task which accesses automatic responses and understands natural verbal relations (Barnes-Holmes, 2010). IRAP requires the subject to combine the target stimuli (smoker, non smoker) and the positive or negative responses to attain the desired consistent or inconsistent response.

Within this research, the data were coded by classifying “non-smokers are cool” or “smokers are lame” in the consistent block. The inconsistent block would be “non-smokers are lame” or “smokers are independent.” Another questionnaire was administered post- IRAP to analyse explicit attitudes toward smokers and non-smokers.

Table 1 Consistent and Inconsistent Stimuli Presented during the IRAP.

Positive Stimuli

Negative stimuli

Cool

Lame

Independent

Boring

Respected

Needy

Attractive

Loner

Fun

Weak

Popular

Loser

Design

A 2 x 2 mixed method design was utilised within this research. The DV’s were responses times on the IRAP and responses in the explicit thermometer scale (Explicits scale). The IV’s in the study were smoking history (smoker/non smoker) and one’s smoking behavioural history (consistent or inconsistent).

Procedure

The researcher first administered the first questionnaire which primed attitudes of smokers and non-smokers. The participants were asked to read the IRAP instructions where they had to answer either Pro-smoker/anti non-smoker or pro non-smoker/anti-smoker. The study required participants to complete two practise blocks and six test blocks. If one is unable to complete these practise blocks, IRAP gives the participant the opportunity to have four more tries, after this they were thanked for their participation and their data is disregarded. IRAP outlines that participants must score over eighty per cent and respond to each trial in less than two seconds. Participants respond by indicating true (pressing the key D) or false (pressing the key K) which interchange during the task. The researcher asked the participant to slow down in the practise blocks if they are making a lot of errors, so to understand the main principal of the research.

It was necessary to have strict lab conditions as the task required complete attention, where possible all distractions were eliminated for example noise, mobile phone etc. IRAP compares the response speed from the consistent and inconsistent blocks. If a positive difference is attained one is more likely to express a pro non-smoker attitude, while if the result is negative one is more associated with expressing a pro-smoker attitude.

The Explicit questionnaire gathered information on the participant’s explicit attitude toward smokers and non smokers on a 0-100 attitude thermometer scale.

Results

ANOVA Implicit Relational Assessment Procedure

A Two way 2 (smoking history: Smoker or Non-smoker) X 2 (Behavioural: Consistent vs. Inconsistent) mixed ANOVA was conducted on the IRAP scores.

Within Subjects ANOVA

There is a significant linear trend on the IRAP scale F(1, 31)= 38.503, p< 0.001, η2=0.554, over the mean values for each level of the factor. The interaction between the smokers and non-smokers and the explicit scale is not significant, F(1, 31)= 0.447, p= .509, η2= .014 over the mean values for each level of the factor.

Between Subjects ANOVA

The main effect whether the smoker versus the non-smoker scores differed on the IRAP was significant F(1, 31) =7.408, p= .011, η2 = 0.193. The interaction between smokers and non smokers social identity preferences was not significant F(1, 31) = 3.413, p=.074, η2 =0.099.

Independent T test

Smokers scored lower (mean= -0.37) than non-smoker’s (mean= -0.14) in the Smoker_D IRAP. The mean difference between conditions was -0.23 and the confidence interval for the estimated population mean difference is between -4.01 and -0.57. An independent t-test showed a difference between conditions was significant (t=2.70, df=31, p=.011 two tailed.)

Smokers scored lower (mean= -0.37) than non-smoker’s (mean= -0.14) in the Non-smoker_D IRAP. The mean difference between conditions was -0.23 and the confidence interval for the estimated population mean difference is between -4.01 and -0.57. An independent t-test showed a difference between conditions was significant (t=2.70, df=31, p=.011 two tailed.)

Paired T-test

The average time it took for smokers to respond to Smoker_D was shorter than of the non-smoker_D trials. The mean difference between conditions was -0.415ms and the 95% confidence interval for the estimated population mean difference is between 0.586 and 0.254. A paired t-test showed that the difference between conditions was significant (t= -5.162, df= 16, p<.001, two tailed.

The average time it took for non-smokers to respond to Non-smoker_D was shorter than of the smoker_D trials. The mean difference between conditions was -.33385ms and the 95% confidence interval for the estimated population mean difference is between -0.539 and -0.128. A paired t-test showed that the difference between conditions was significant (t= 4.36, df= 16, p<.001, two tailed.

ANOVA of Explicit Preferences toward Smoking-Related social identity Scale

A Two way 2 (smoking history: Smoker or Non-smoker) X 2 (Behavioural: Consistent vs. Inconsistent) mixed ANOVA was conducted on the Explicit preferences toward smoking related social identity ratings.

Within Subjects ANOVA on the Explicit Scale

There is a significant linear trend on the explicit scale, F(1, 30)= 7.952, p= .008 η2 = .210, over the mean values for each level of the factor. There is also a significant interaction between the smokers and non-smokers and the explicit scale, F(1, 30)= 7.522, p= .007, η2= .220, over the mean values for each level of the factor.

Between Subjects ANOVA on the Explicit Scale

The main effect whether smoker versus the non-smoker scores differed on the Explicit scale was significant F(1, 30) = 13.952, p=.001, η2 = 0.317. There was a significant interaction between non-smokers and smokers on the explicit’s scale F(1, 30) = 12.506, p=0.001, η2 = .294.

Independent T-test for Explicit Scale

Smokers scored positively (mean= -0.696) than non-smoker’s who responded negatively (mean= 0.711) in the Explicit Smoker Scale. The mean difference between conditions was 1.407 and the confidence interval for the estimated population mean difference is between -2.20 and -0.605. An independent t-test showed a difference between conditions was significant (t=-3.586, df=30, p=.001 two tailed.)

Smokers scored less positively (mean= .657) in comparison to the non-smoker’s(mean= 0.690) in the Explicit Non-Smoker Scale. The mean difference between conditions was -0.33 and the confidence interval for the estimated population mean difference is between -3.99 and -0.333. An independent t-test showed a difference between conditions was not significant (t=–0.185, df=30, p=.855 two tailed.)

Paired T-test

Smokers responded more positively to the Explicit Smoker Scale (-0.696) compared to the Explicit Non-Smoker Scale (0.6578). The mean difference between conditions was 1.353 and the 95% confidence interval for the estimated population mean difference is between 0.695 and 2.01. A paired t-test showed that the difference between conditions was significant (t= 4.36, df= 16, p<.001, two tailed).

Non-smokers responded positively toward the Explicit Non-smoker scale (0.6909), however they scored extremely negative toward the Explicit Smoker scale (0.711). The mean difference between the conditions was -0.02 and the 95% confidence interval for the estimated population mean difference was between -0.729 and 0.752. A paired t-test showed that the difference between conditions was not significant (t= -0.056, df= 14, p=0.956, two tailed),

Correlation between IRAP and Explicit Attitude Scale

Pearson’s Correlation of Non Smoker D (IRAP) and Explicit Non-Smoker

Non-Smoker_D

Explicit Non Smoker

Pearson Correlation

-.008

Sig (2-tailed)

.965

Pearson’s Correlation of Smoker_D (IRAP) and Explicit Smoker

Explicit Smoker

Pearson’s Correlation

Sig (2-tailed)

Overall there was a significant difference between smokers and non smokers on the IRAP in the consistent and inconsistent variable; however the effect size was relatively small. Smokers were found to be positive toward smokers and non-smokers both on the IRAP and explicit’s scale. Non-smokers were consistent holding positive regard of themselves. There was a difference in the non-smokers implicit and explicit responses towards smokers, their explicit responses indicate negative regard for smokers, yet implicitly they were ambiguous. There was an interaction effect in the explicit scale as non-smokers were more negative toward smokers than positive toward non-smokers. There was no significant relationship between responses on the explicit smoker scale and implicit smoker_D scale. Likewise no significant correlation was evident between the explicit non-smoker scale and the non-smoker_D scale.

FIGURE 1; Responses on IRAP demonstrating preferences toward smoking-social identity.

FIGURE 2: Smokers and Non-Smokers responses on the Explicit Thermometer scale,

Discussion

Within this research, we fail to accept our null hypothesis, as smokers regard themselves and non-smokers positively both explicitly and implicitly, which is consistent with the findings of Vahey et al. Within our second hypothesis, we accept parts; it was found that non-smokers will score higher in the pro-non smoker condition, as it is relevant to their behavioural history and social identity. However the research demonstrates that non-smokers will explicitly express negative attitudes toward smokers, yet implicitly respond in an ambiguous manner. The IRAP methodology is proficient as it reveals social buffering prevalent within non-smokers. The interaction effect indicated that non-smoker were explicitly marginally more negative toward smokers than positive toward non-smoker’s. This may be due to the more prevalent negative social attitude toward smokers, while non-smokers receive less positive reinforcement.

In previous research non-smoker adolescents did not implicitly or explicitly condone smokers due to peer pressure as this behaviour is considered cool (Vahey et al, 2010). Our results differ as the adult population are more influenced by the negative stereotype of smokers which is manifested by the governmental laws, advertising etc. People are more likely to express this negative opinion as it is socially desirable however they may not inherently agree.

One of the strengths of this research is that implicit attitudes are exposed concerning smoking-related identities, which increases the relevance of IRAP. For instance this research identifies social buffering as non-smokers conflict in their implicit and explicit attitudes toward smokers.

There are a number of shortcomings within this research for example there is a small effect size, which reduces the generalisability of this research to the general population. This study was not carried out in a lab setting which could have affected people’s responses. For instance, there may have been more distractions prevalent which could have delayed people’s responses on the IRAP in the consistent and inconsistent condition. An alternative shortcoming to be considered is that some of the sample may have been social smokers and may not have taken the smoker-identity.

References

Barnes-Holmes, D., Barnes-Holmes, Y., Stewart, I., & Boles S., (2010) A Sketch of the Implicit Relational Assessment Procedure (IRAP) and the Relational Elaboration and Coherence (REC) Model. The Psychological Record, (60) p. 527-542

Dowling, K., (January 18, 2010) “Smoking: Ireland’s ban re-examined” Retrieved from http://www.euranet.eu/eng/Archive/News/English/2010/January/Smoking-Ireland-s-ban-re-examined (accessed on the 15/10/2010).

Vahey, N., Boles, S., & Barnes-Holmes, D., (2010) “Measuring Adolescents’ Smoking-related Social Identity Preferences with the Implicit Relational Assessment Procedure (IRAP) for the First Time: A Starting Point that Explains Later IRAP Evolutions” National University of Ireland Maynooth and Oregan Research Institute p. 1-45

Appendix 1

Analysis of Variance of Implicit Relational Assessment Procedure (IRAP)

Tests of Within-Subjects Effects

Within Subjects ANOVA

df

F

Sig.

IRAP

Sphericity Assumed

1

38.503

.000

Greenhouse-Geisser

1.000

38.503

.000

Huynh-Feldt

1.000

38.503

.000

Lower-bound

1.000

38.503

.000

IRAP interaction with Smoker Non-smoker

Sphericity Assumed

1

.447

.509

Greenhouse-Geisser

1.000

.447

.509

Huynh-Feldt

1.000

.447

.509

Lower-bound

1.000

.447

.509

Error(IRAP)

Sphericity Assumed

31

Greenhouse-Geisser

31.000

Huynh-Feldt

31.000

Lower-bound

31.000

Tests of Between-Subjects Effects in IRAP

Df

F

Sig.

Intercept

1

3.413

.074

Smoker/Non-smoker

1

7.408

.011

Error

31

Appendix 2

Analysis of Variance of Explicit Scale

Tests of Within Subject Effects

Within Subjects ANOVA

Df

F

Sig.

Explicit Scale

Sphericity Assumed

1

7.952

.008

Greenhouse-Geisser

1.000

7.952

.008

Huynh-Feldt

1.000

7.952

.008

Lower-bound

1.000

7.952

.008

Explicit Scale interaction with Smoker Non-smoker

Sphericity Assumed

1

.891

.007

Greenhouse-Geisser

1.000

.891

.007

Huynh-Feldt

1.000

.891

.007

Lower-bound

1.000

.891

.007

Error(Explicit)

Sphericity Assumed

30

Greenhouse-Geisser

30.000

Huynh-Feldt

30.000

Tests of Between-Subjects Effects in IRAP

Df

F

Sig.

Intercept

1

12.506

.001

Smoker/Non-smoker

1

13.952

.001

Error

30

Appendix 3

Independent T-test for IRAP, tests differences on Smoker_D and Non-Smoker_D

95% Confidence Interval of the Difference

T

Df

Sig (2-tailed)

Mean Difference

Std Error Diff

Lower

Non Smoker_D

Equal variances assumed

-1.513

30

.141

-15451

.10214

-3.6310

Smoker_D

Equal variances assumed

-2.684

30

.012

-.23642

.08809

-.41631

95% Confidence Interval of the Difference

t

Df

Sig (2-tailed)

Mean Difference

Std Error Diff

Lower

Explicit Smoker

Equal variances assumed

-3.586

30

.001

-1.40719

.39243

-2.20865

Explicit Non-Smoker

Equal variances assumed

-.185

30

.855

-.03316

.17937

-.39947

Appendix 4

Independent T-test for Explicit Scale, tests differences between Explicit Smoker and Explicit Non-Smoker.

Appendix 5

Paired Samples T-Test for Smokers in IRAP

Descriptives for Smoker in IRAP

Mean

N

Std. Deviation

Std. Error Mean

Pair 1

Smoker_D

-.3726

17

.22809

.05532

Non_Smoker_D

.0432

17

.26499

.06427

Paired Differences

t

df

Sig. (2-tailed)

Mean

Std. Deviation

Std. Error Mean

95% Confidence Interval of the Difference

Lower

Upper

Smoker_D – Non_Smoker_D

-.41575

.33209

.08054

-.58650

-.24500

-5.162

16

Paired Sample T-test for Smokers in IRAP

Paired Samples T-Test for Non-Smokers in IRAP

Descriptives for Non-Smokers in IRAP

Mean

N

Std. Deviation

Pair 1

Smoker_D

-1.362

15

.27026

Non_Smoker_D

.1977

15

.31286

Paired Samples T-Test for Non-Smokers in IRAP

Paired Differences

t

df

Mean

Std. Deviation

Std. Error Mean

95% Confidence Interval of the Difference

Lower

Upper

Smoker_D – Non_Smoker_D

-.33385

.37074

.09572

-.53916

-.12854

-3.488

14

.004